1Virome characterization of Cryphonectria parasitica isolates from Azerbaijan unveiled a new

2mymonavirus and a putative new RNA virus with a GDD motif unrelated to existing viral
3sequences

4Forgia, M., Isgenderli, E., Aghayeva, D., Turina, M.

5 Introduction

6Cryphonectria parasitica is the ascomycete causal agent of chestnut blight disease that endangered
7chestnut cultivation in Europe and America during the 20 th century (Dawe and Nuss, 2013). In
8Europe, chestnut blight disease control was achieved through spreading of the Cryphonectria
9hypovirus 1 (CHV1) that cause hypovirulence in the fungal host (Grente, 1965; Milgroom and
10Cortesi, 2004). After the production of an efficient system to build a CHV1 infectious clone and the
11availability of protocols for protoplast production and genomic transformation, C. parasitica
12became an important model to study hypovirulence and host-mycovirus interaction (Nuss, 2005)
13despite a relatively low number of mycovirus infecting C. parasitica have been described so far.
14Indeed, even if Next Generation Sequencing (NGS) approaches have been shown to represent the
15most valuable tool to deeply investigate the viral diversity in a collection of fungal isolates (Nerva,
162016) there are no works yet implementing these techniques on C. parasitica reported in literature
17and most of the mycoviruses infecting this model fungus have been identified through dsRNA
18isolation (Hillman and Suzuki, 2004). Mycoviruses infecting C. parasitica show different genome
19organization and have dsRNA genomes or positive ssRNA genomes. Taxonomically they belong to
20different families: Hypoviridae, Reoviridae and Narnaviridae; so far, no negative stranded
21mycovirus have been isolated from C. parasitica (Eusebio-Cope et al., 2015) . Moreover,
22different works showed the possibility to infect C. parasitica with different mycoviruses through
23purified viral particle transfection allowing host-virus interaction studies even with viruses not
24originally isolated from the model fungal host (Nerva et al., 2017). During recent years C.
25parasitica has been reported in Azerbaijan where chestnut cultivation occurs (Aghayeva and
26Harrington, 2007). C. parasitica isolates collected in Azerbaijan shown a high genetic homology
27compared to those from Georgia. In facts, almost all, but one Azerbaijani genotypes are present in
28Georgia and belong to the Georgian gene pool of C. parasitica (Aghayeva et al., 2017; Prospero et
29al., 2013). To date, little is known about the dissemination of CHV1 through Azerbaijani population
30of C.parasitica; aim of this work was to collect a large number of fungal isolates from Azerbaijani
31population of C. parasitica and describe through a high throughput approach the virome of the
32collected isolates in order to find evidence of CHV1 infections and to describe possible new

1 1
2
33mycoviruses infecting this fungal host. CHV1 detection in Azerbaijan’s isolates could give
34information about the possibility to control chestnut blight disease though spreading hypovirulent
35isolates; furthermore, the discovery of new viruses infecting C. parasitica as natural host could
36allow investigating the host-virus molecular mechanisms thanks to the possibility allowed by this
37genetically treatable fungal model.

38 Results

39Isolates collection and identification

40Isolation of C. parasitica was performed in three (Qabala, Shaki, Zagatala) districts (Figure 1). Bark
41samples were collected from a total of 77 cankers and from 58 of them isolated in pure culture.
42Most of the remaining bark samples also yielded C. parasitica colonies, but these were heavily
43contaminated with other fungal species and were not included in further analyses. Some variation
44was observed among the isolates with respect to their culture morphology (i.e., pigmentation and
45asexual sporulation) on PDA. About 37 culture showed an orange and 21 whitish-orangephenotype
46on PDA 10 days post inoculation, but white cultures, typical for C. parasitica-infected with CHV-1
47were not observed. Incidence of sexual reproduction was estimated observing C. parasitica
48stromata under the microscope. Both perithecia an picnidia were observed in most of examined
49stromata. All examined perithecia contained asci with ascospores. Perithecia were present in all
50districts with the highest prevalence found in Oghuz and Ismailli (43.8%). Both mating types of the
51fungus were found in all three districts. From the isolates collected, fifty-five were chosen for
52virome exploration through NGS approach. A list of isolates with information about collection site
53and date can be found in Supp. table 1.
54Virome exploration

55Fifty-five isolates from the Azerbaijani C. parasitica collection were inoculated in liquid cultures to
56obtain a high amount of mycelia. From the obtained mycelia, total RNAs were extracted and pooled
57in two samples for illumina sequencing; pooled samples composition is showed in Supp. table 2.
58After ribosomal RNA depletion, library construction and illumina sequencing, the two libraries
59obtained were assembled and viral contig were retrieved through a blast-based bioinformatic
60pipeline. From this analysis, we were able to identify one virus related to the Mymonaviridae family
61named Cryphonectria parasitica mymonavirus 1 (CpMyV1) and one ORFan sequence with peculiar
62feature suggesting a viral origin.

63CpMyV1 characterization

3 2
4
64After blasting the transcriptomes against our custom viral database, we identified one 6439 bp long
65contig encoding for three major ORF in NGS sample 1 named Contig1. The third ORF was partial
66and showed a Mononegavirales RdRP domain with a close homology to RdRP encoded by the
67sclerotimonavirus Soybean leaf-associated negative-stranded RNA virus 2. Quantitative RT-PCR
68analysis on cDNAs performed on each of the fungi hosted in the NGS Sample 1 revealed that
69Contig1 was present only in the isolate ACP 28 (NGS Sample1, Supp. table 3) and that no
70amplification was observed using the same qRT-PCR primer on ACP 28 DNA Data not shown);
71this result confirmed the viral nature of Contig1 excluding the possibility that the virus is integrated
72in the fungal genome. Viruses belonging to Mymonaviridae family have a negative single-stranded
73RNA genome and encode for an ORF in the 5’ extremity downstream the ORF encoding the RdRP.
74To complete the sequence of the C. parasitica viral contig, we retrieved the 5’ terminal ORF from
75Soybean leaf-associated negative-stranded RNA virus 2 (one of the viruses showing higher
76homology to our contig) and we blasted it against our transcriptome from NGS Sample 1. We
77identify one 2901bp contig named Contig2 hosting the remaining part of the genome encoding the
78mymonavirus RdRP and one ORF in the 5’ part of the contig. Since Contig1 and Contig2 did not
79overlap, we designed a primer couple to amplify the missing part to complete the RdRP coding
80sequence. After PCR amplification, cloning of the PCR product and sequencing of 6 different
81colonies we obtained the consensus sequence of the missing part of the viral genome between
82Contig1 and Contig2 and we named the virus Cryphonectria parasitica mymonavirus 1 (CpMyV1).
83Rapid Amplification of cDNA ends (RACE) on the 5’ and 3’ extremity of the viral genome were
84performed and results added just two nucleotide at the 5’ extremity and one nucleotides at the 3’
85extremity. Furthermore, PCR amplification of 1kb overlapping fragments of the viral genome was
86performed in order to cover most of the viral sequence and to confirm the assembly obtained
87through bioinformatic approach; as shown in Suppl. figure1 previous results were confirmed by
88PCR analysis with ten different primer couples listed in Suppl. table 3. Taken together these results
89show the genome organization and main features of the first complete sequence of a negative-sense
90single stranded RNA virus infecting C. parasitica (Fig. 2); CpMyV1 is 9318 bp long and it encodes
91for five ORF named ORF 1 to 5; blastp result obtained submitting each ORF against the NCBI
92database are shown in table 1 together with information about reads abundance in the NGS sample
93hosting the virus. ORF1 produces a 293 amino acid long putative protein that shows a
94USP19_linker domain when submitted on NCBI domain finder, blastp analysis shows little
95homology with other putative proteins encoded by other mycoviruses belonging to Mymonaviridae
96family. ORF2 encode for what is likely the nucleoprotein of CpMyV1, a predicted protein 391
97amino acid long showing close homology to other putative nucleoprotein encoded by

5 3
6
 98mymonaviruses; the predicted protein doesn’t show any conserved domain on NCBI domain finder.
99ORF3 is a 42 amino acid protein without homology to known protein. ORF4 encodes for the viral
100RdRP, the protein predicted is 1931 amino acid long and shows a Mononegavirales RdRP domain
101together with a Mononegavirales mRNA capping domain. The last putative protein, produced by
102ORF5, is a 178 amino acid long sequence showing homology with ORFs encoded from the 5’
103extremity of several mycoviruses belonging to Mymonaviridae family, no conserved domains are
104detected with NCBI domain finder. A typical feature showed by mymonaviruses is the conserved
105sequence between ORFs and between the last ORF and 5’UTR extremity, by aligning these
106sequences retrieved from CpMyV1 we were able to observe a conserved nucleotide patterns that is
107shown in figure2 (panel b). Phylogenetic analysis performed on the RdRP sequence (figure3) placed
108CpMyV1 in the Mononegavirales order and in the Mymonaviridae family. CpMyV1 is likely a
109member of the only genus contained in the Mymonaviridae family called Sclerotimonavirus that
110hosts only mycoviruses; indeed, viruses included in the genus are highlighted in the tree and are
111found in the same monophyletic branch that is well supported by the bootstrap analysis.

112Characterization of an ORFan sequence as a putative viral segment

113Since a major limit of homology based method to find viruses is the loss of viral contig with low or
114no homology to known sequences, we filtered the transcriptomes to remove the fungal transcript
115and retrieve only the contigs showing “no hit” result to any sequence contained in the NCBI nr
116database. The same procedure was performed on a pooled transcriptome we used to describe
117Rhizoctonia solani virome in a previous work (Picarelli et al, 2019). After this analysis, we removed
118the smallest transcripts and selected contigs encoding for at least a 10 KDa protein; these putative
119proteins were blasted against the same sequences retrieved from the R. solani RNAseq. From this
120analysis, a contig of 4623 bp encoding for 2 putative ORFs was isolated both from NGS sample 1
121and 2 and two contigs of 4868 bp and 3729 bp were isolated from R. solani NGS sample. The three
122contigs were provisionally named Cryphonectria parasitica ambivirus 1 (CpaV1), and Rhizoctonia
123solani ambivirus 1 and 2 (RsaV1 and 2). A schematic representation of the genome organization of
124CpaV1 with typical ambisense orientation of two ORFs encoding sequences can be observed in
125Figure4). These contigs show similar ORFs pattern and conserved domains among homologous
126proteins (Figure 4 panel b and c) and since in all cases one of the two predicted proteins contains a
127GDD domain, the hallmark of RNA-dependent RNA-polymerase, further in-vivo characterizations
128of these genome organization have been performed. To understand how CpaV1 is spread across the
129fungal isolates we designed qRT-PCR primers (Suppl. table 3) and we performed qRT-PCR
130analysis on all the C. parasitica isolates cDNA; results (Supplementary table S2) show that twenty

7 4
8
131two isolates from the initial fifty five host CPaV1, for further characterization we used as a
132reference isolates ACP 34, 36 and 43. Since no signal was detected in qRT-PCR reactions
133performed on DNA extracted by infected C. parasitica we can assume that CpaV1 is an RNA
134sequence not integrated in the fungal genome (Crouch et al., 2020); further confirmation of this
135evidence can be observed in Suppl. figure2 were PCR reactions to amplify three hundred bp long
136sequences from both ORFA and ORFB coding sequence were performed. As shown by the figure,
137PCR amplification was obtained only in the sample containing cDNAs as a template and no
138amplification was observed in the samples containing ACP 34, 36 and 43 DNAs. PCR fragments
139obtained from the analysis were cloned and sequenced to use them as northern blot probes: position
140of the probes and their orientation is shown in Figure 4 a. Remarkably even if the two ORFs
141showed conservation compared to contigs isolated from different fungi, they had different
142disposition on the sequences: the two ORFs are always disposed in an ambisense orientation and
143they could be oriented outward (pointing to the 5’ and 3’ UTRs as observed in C. parasitica orfan
144contig) or inward, toward the center of the contig (as observed in the two contigs from R. solani).
145This evidence brought to the hypothesis that these contigs could derive from a circular sequence,
146thus we designed PCR primers amplifying upstream the 5’ end and downstream the 3’ end
147obtaining a band that was cloned and sequenced showing the possibility to reconstruct a circular
148sequence from the initial contig and allowing us to confirm that the sequence is complete. To
149further confirm that CpaV1 sequence is complete, we amplified 1kbp long overlapping PCR
150fragments covering the entire sequence and result showed confirmed the expected size of each
151portion (Supplementary Figure S3). As shown in figure4 the two ORF (named ORFA and ORFB)
152have an ambisense orientation, putative proteins produced by CpaV1 were again submitted to a
153blastp search against the NCBI database and results obtained are shown in table1 together with
154information about reads mapping on the nucleotide sequence for both NGS samples. ORFA is a 757
155amino acid long protein, it doesn’t show any conserved domain and blastp analysis show homology
156with a partial protein encoded by an ORFan sequence isolated from Agaricus bisporus and that
157belongs to an RNA sequence not integrated in the genome(Deakin et al., 2017). The predicted
158protein hosts a GDD domain that is considered as the hallmark of the viral replicases (GDD
159domain citation, cite A.bis work). ORFB encodes for a 663 amino acid sequence that has
160homology with hypothetical proteins from several fungi; these proteins could derive from
161integration events happened between host fungi and other CpaV1-like sequences. Even if initially
162this sequence have been isolated through searching between the “no hit” contigs, blast hits have
163been obtained in this analysis because the evalue cutoff set on NCBI blast is higher compared to
164evalue set on DIAMOND blast (see method section for further details). Northern blot analysis were

9 5
10
165performed using fragments amplified on both ORFs in each sense to detect positive and negative
166sense RNA; analysis was carried on using ACP 34 and 36 as positive isolates and ACP 23 as
167negative (see figure 4, panel a, for information about northern blot probes location on the sequence).
168Results showed two bands specific to the infected isolates that are present with all the probe tested
169even directed to ORFA or ORFB (Figure5); low exposition time allow to appreciate that the lower
170of the two bands appear to be less intense than the higher and probes detecting the negative sense
171RNA give a less intense signal for all ORFs. ACP 23 isolate included as negative control showed
172few bands that are likely aspecific signal coming from probe-ribosomal RNA interaction (confirmed
173overlapping with methylene blue stained membrane). Comparing these bands to the one obtained by
174hybridization with a Tomato brown rugose fruit virus (TBRFV) infected tomato that was added on
175the RNA gel as a marker(Salem et al., 2016), we could estimate that the lower band should be
176compatible with CpaV1 size and we hypothesize that the higher band could derive from a structure
177assumed by the circular entity of the genome causing a difference in the electrophoresis migration
178or to a possible dimer. To understand if CpORFan really has a circular sequence we treated
179extracted RNA from ACP 34 with RNAse R enzyme; RNAse R is an exonuclease working
180specifically on linear RNAs and circular RNA molecules should be naturally protected from his
181action. RNA from ACP 34 and TBRFV infected tomato were treated with RNAse R for thirty
182minutes, one or two hours, then, the RNAs were migrated and blotted to perform northern blot
183analysis. Figure6 shows results obtained from this experiment, where RNAse R action was
184confirmed because ribosomal RNAs were not visible anymore even after thirty minutes treatment
185but they were still present in the samples just mixed with RNAse R reaction buffer. Both CpaV1
186and TBRFV, used as negative control, didn’t give any signal after hybridization with specific
187northern blot probes suggesting that CpaV1 is probably able to produce dimers that allow PCR
188amplification of fragments overlapping 5’ and 3’ ends and reads mapping across this area. This
189result also tells that the higher band detected in northern blots is probably a dimer formed by two
190CpaV1 molecules. Looking at northern blot results, it appears that the most abundant form of
191CpORFan is the one where ORFA is found in positive orientation, with the higher band (probably
192associated to the dimer form) more intense than the lower one; thus the dimer is likely the most
193present form in the host.

194 Discussion

195Cryphonectria hypovirus 1 is found across Europe, in China and in some parts of North America;
196where the dissemination was less efficient because of a higher genetic diversity of C. parasitica
197population. CHV1 was also detected in Georgia (https://www.mdpi.com/1999-4915/10/12/687)

11 6
12
198which is likely the source of Azerbaijani C. parasitica isolates. After CHV1, three other
199hypoviruses have been reported to infect C. parasitia: Cryphonectria hypovirus 2, 3 and 4. All these
200hypoviruses are able to induce hypovirulence, even if most of the works are concentrated on CHV1
201because of the easy manipulation and the better performances as biocontrol agent. CHV2,3 and 4
202have all been found in North America, but CHV2 was also detected in China, in both cases the areas
203of diffusion are limited. Using high throughput approaches we demonstrated that our collection of
204C. parasitica isolates from Azerbaijan doesn’t host Cryphonectria hypovirus 1 (or any of the other
205Hypoviruses found in other populations) and that spreading of the hypovirulent strain seems to be
206impossible at least starting from a CHV1 infected isolate already present in the environment.
207Moreover this work unveiled Cryphonectria parasitica mymonavirus 1, the first negative-sense
208single stranded RNA virus infecting C. parasitica as a natural host; this result could be important to
209characterize the interaction between these viruses and the host thanks to the possibility allowed by
210the fungal model; indeed in some cases viruses belonging to the Mymonaviridae family and the
211Sclerotimonavirus genus have been associated to hypovirulence (Liu et al., 2014), so further
212characterization could be made trying to obtain isogenic isolates not infected with CpMyV1 to
213observe any difference in the pathogenicity. The major limit of homology-based method to find
214viruses is the loss of viral contig with low or no homology to known sequences. Recently, putative
215nucleocapsids from Bunya-like mycoviruses have been found in NGS experiments by comparing
216the contigs without a hit in the NCBI database from two NGS experiments containing two different
217Bunya-like mycoviruses and retrieving the nucleocapsid contigs thanks to the homology between
218the two protein sequences (Nerva et al., 2019). We decided to perform the same experiment on C.
219parasitica NGS transcriptome comparing it to the transcriptome obtained from Rhizoctonia solani
220RNAseq from a previous work (Picarelli et al., 2019b). Results allowed us to identify an ORFan
221contig encoding for two ORFs in ambisense orientation, the sequence belongs to an RNA molecule
222that is not integrated in the host genome therefore supporting its viral origin (self-replicating RNA
223molecule); one possibility is that another RdRP replicates in trans this molecule, but our screen of
224C. parasitica virome allows us to exclude the occurrence of other viral RdRPs in the samples. The
225presence of a GDD domain in the amino acid sequence from putative ORFA can suggest that the
226ORFan contig we dected could be a virus, and northern blot analysis together with PCR analysis
227showed that dimer formation could be possible for CpaV1. CpaV1 has homology with contigs
228isolated from a different fungal host, but ORF orientation is not conserved between these
229sequences; a possible explanation could be that the ability to form dimer could allow mistakes in the
230bioinformatic reconstruction of the ambivirus ORFan genome or that these dimers could divide
231keeping the two ORFs but displaying the opposite orientation.

13 7
14
232Abundant dimer accumulation during replication of negative strand RNA virus was previously
233reported (Bertran et al., 2016) and hypothesized to be linked to a regulation of replication similar to
234that caused by accumulation of defective interfering RNA (DI-RNA).

235Notably, this is not the first time that sequences like CpaV1, RsaV1 and 2 are identified while
236exploring the virome of a fungus. During the characterization of the virome from a collection of
237isolates of Agaricus bisporus, several ORFan belonging to RNA sequences not integrated in the
238fungal genome were collected thanks to the different bioinformatic approach used for that study.
239Between these, ORFan 1 display an organization that is possible to connect to a dimer version of a
240CpaV1-like sequence with a mis-assembly that leaves ORFA uncomplete. Nevertheless, partial
241sequence of ORFA putative protein (at the 5’ extremity of the contig) shows conserved motifs when
242aligned to ORFA from CpaV1, RsaV1 and 2 (Figure 4, panel b and c). This is again an evidence
243suggesting the possible viral nature of these sequences that are found in several different fungi.

244

245Methods

246Chestnut blight detection and sampling

247A total of three chestnut growing sites in three districts in North-Eastern Azerbaijan were surveyed
248in December 2018 and April 2019 (Figure 1). Chestnut trees were visually inspected for symptoms
249of C. parasitica, such as bark cankers on stems or branches, wilting leaves on individual branches
250in the crown (Rigling and Prospero, 2017). Bark pieces (ca. 2 × 2 cm) with visible C. parasitica
251stromata were removed with a knife from the cankers. Each time the knife was dipped in 70%
252ethanol and flamed after picking canker. To avoid the isolation of clones of the same genotype, only
253one single canker per tree was sampled (Milgroom and Lipari, 1995). In each district, 25–30
254cankers were sampled.
255
256Fungal isolation, recognition and maintenance
257Bark samples were examined under a dissecting microscope for the presence of pycnidia or
258perithecia of the fungus (Prospero et al., 2013) . Perithecia were dissected from the stromata and
259crushed in a drop of water to verify presence of asci with ascospores. Stromatal tissue of C.
260parasitica was removed from the bark samples with a sterile needle under a dissecting microscope
261and placed on potato dextrose agar (PDA: 39 g/L, Difco™, Becton, MD, USA) for isolation. The
262PDA plates were incubated at room temperature, and one outgrowing culture (isolate) per bark
263sample was kept for further analyses. Fungal isolates were maintained on 60 mm diameter petri dish

15 8
16
264containing PDA media at 26 °C with a light/dark cycle of 12 hours. To obtain large amount of
265fungal biomass for molecular analysis, fungi were inoculated in 50 ml of liquid EP complete media
266(Puhalla and Anagnostakis, 1971). After 3 days, liquid cultures were harvested with a Buchner filter
267covered with miracloth (Merck, Darmstadt, Germany) in a vacuum pump to separate mycelia from
268the liquid media. Obtained mycelia was lyophilized and kept at -20 °C.
269
270RNA and DNA extraction
271RNA extraction was performed using plant total spectrum RNA (Sigma-Aldrich, St. Louis, MO,
272USA) starting from 30 mg of lyophilized myceliaas previously described (Nerva et al., 2018). DNA
273was extracted starting from 50 mg of lyophilized mycelia in a 2 ml O-ring tube with 0.5 mm
274diameter glass beads. Mycelia was broken using a Fast prep-24 (MP Biomedicals, Irvine, CA,
275USA), adding 700 µl of 2X STE/SDS and 700 µl of phenol pH 8.0 and repeating the fast prep step.
276Tubes were then centrifuged for 3 minutes at max speed in a table centrifuge and upper phase was
277collected in new 2 ml tubes. Obtained phases were washed twice with equal volume of 24:1
278chloroform: isoamyl alcohol and upper phases were collected every time. Nucleic acids were then
279precipitated with 2 volumes of 100% ethanol, 40% of the phase volume of 3 M sodium acetate pH
2803.5 and keeping the tubes at -80 °C for 20 minutes. Tubes were centrifuged 3 minutes at max speed
281and supernatants were discarded, the pellets were resuspended in 50 µl of sterile water and
282quantified with a Nanodrop 2000 (Thermo-Fisher scientific, Waltham, MA, USA). Obtained DNAs
283were diluted to 10 ng/µl for PCR use.

284Illumina sequencing and bioinformatic analysis

285Total RNAs from each isolates of the collection of C. parasitica were pooled in two samples
286maintaining an equal proportion of RNA coming from each isolate. Composition of the two pooled
287samples is shown in Suppl. table2. Samples were then sent to Macrogen (Seul, South Korea) for
288ribosomal RNA depletion, library construction and Illumina sequencing using TruSeq Stranded
289Total RNA LT Sample Prep Kit (Gold) kit (Illumina, San Diego, CA, USA) and producing 100
290million reads of 100 bp length in paired end. Reads were cleaned using bbmap (cite) functions in
291order to eliminate short reads, low quality reads and artifacts. Cleaned reads were assembled using
292Trinity version 2.3 (Haas et al., 2013) and obtained transcriptome were blasted against a custom
293viral database (https://osf.io/c9x2p/) using DIAMOND blastx function (Buchfink et al., 2015).
294Blast result was manually checked to retrieve the viral contigs and to further characterize them in
295vivo. Mapping reads on the viral genomes were obtained producing an alignment in sam format
296with bowtie2 (Langmead and Salzberg, 2012); sam file was converted in bam format with samtools

17 9
18
297view and sorted with samtools sort (Li et al., 2009). Using samtools index a sort.bam.bai file was
298generated, and result was displayed on Tablet(Milne et al., 2016). ORF prediction was performed
299using ORF finder NCBI tool and conserved protein domains were searched through submission on
300NCBI conserved domain search. Bioinformatic pipeline adopted is well described in our previous
301publication (Nerva, 2018) and custom scripts used for the analysis are available at the following
302repository (scripts links still to add on OSF repo)

303qRT-PCR analysis
304To associate each viral contig to the correct fungal isolate contained in the pooled sample, qRT-
305PCR analysis was performed designing specific primers on each viral contig using Primer3
306(Untergasser et al., 2012) allowing an amplification size between 75 and 120 nucleotides. qRT-PCR
307were performed on a CFX connect (Bio-rad, Hercules, CA, USA) apparatus and each reaction was
308prepared in 10 µl using the same protocol and reagents as described in previous publication
309(Picarelli et al., 2019a).
310
311PCR analysis, cloning and northern blot probes production
312RT-PCR was performed to obtain cDNAs from infected isolates to produce Northern blot probe for
313CpaV1 and to obtain the missing part of the viral genome for Cryphonectria parasitica
314mymonavirus 1. cDNAs were synthesized from the extracted RNAs using high capacity cDNA kit
315(Thermo-Fisher scientific, Waltham, MA, USA), following manufacturer instructions and obtained
316cDNAs were diluted 1:5 in sterile water. RT-PCRs were performed using One Taq DNA
317polymerase (NEB, Ipswich, MA, USA) following manufacturer instruction; bands were isolated
318from agarose gel using Zymo DNA gel recovery kit (Zymo Research, Irvine, CA, USA). Purified
319bands were ligated into pGEMT vector using pGEMT easy vector system (Promega, Madison, WI,
320USA) and transformed in E. coli DH5α. Obtained colony were screened through PCR using M13
321forward and reverse primers with OneTaq DNA polymerase and sent to sanger sequencing to
322Biofab s.r.l (Rome, Italy). A list of primers used for the PCR reactions is shown in Suppl. table 3.
323Plasmids were digested with EcoICRI restriction enzyme (Thermo-Fisher scientific, Waltham, MA,
324USA) as shown by producers. Radio labeled RNA probes were produced using T7 polymerase from
325maxiscript kit (Invitrogen, Carlsbad, CA, USA) following manufacturer instruction and using P 32
326labeled Uridine and as template the EcoICRI-linearized plasmid; after the transcription DNA
327template was removed through TurboDNAse (Invitrogen, Carlsbad, CA, USA) treatment for 20
328minutes at room temperature and probes were cleaned on column using RNA gel recovery kit
329(Zymo Research, Irvine, CA, USA).
330Rapid Amplification of cDNA Ends (RACE)
19 10
20
3315’ and 3’RACE were performed to complete the sequence of Cryphonectria parasitica mymonavirus
3321 following a methodology previously described (Rastgou et al., 2009)(Rastgou et al 2009). From
333total RNA extracted as described, PCR reaction were performed using specific primers amplifying
334CpMyV1 ends: for 5’ RACE a reverse primer annealing specifically 300 bp from the putative end
335was used for cDNA synthesis and for 3’ RACE a forward primer annealing 300 bp from the
336putative end was used for cDNA synthesis(Suppl. table 3). Terminal Deoxynucleotidyl Transferase
337(Promega, Madison, WI, USA) was used as suggested by the manufacturer to add dATP and dGTP
338to the obtained tailed cDNA , half of the reaction was added with dATP and half with dGTP. A
339PCR reaction was then performed using a polyT primer for dATP reactions and polyC primer for
340dGTP reaction together with a primer specific for CpMyV1 and amplifying aroung 250 bp from the
341putative ends ( Suppl. table 3). PCR bands obtained were purified and cloned as already explained.
342To reconstruct the full sequence of CpMyV1 we sequenced clones from dATP and dGTP reactions
343to understand clearly which is the last nucleotide of the viral genome on each end.
344Northern blot analysis
345To perform Norther blot analysis on CpaV1 RNA, total RNA from infected and non-infected fungi
346was separated on Hepes-EDTA agarose gel after denaturation with deionized glyoxal and DMSO as
347described in literature (Ferriol et al., 2018)(. RNA gels were blotted on Immobilon-Ny+ membrane
348(Merck, Darmstadt, Germany) and blots were hybridized with radioactively labelled probes at 60° C
349using ULTRAhyb™ Ultrasensitive Hybridization Buffer (Thermo-Fisher scientific, Waltham, MA,
350USA) following manufacturer instruction.
351DNAse R treatment
352To investigate if CpORFan genome is circular RNA we treated extracted RNAs with RNAse R
353(brand). We set the reaction putting 15 µg of ACP 34 RNA in a 20 µl reaction containing 2 µl of
35410X reaction buffer for RNAse R and collecting 5 µl of the mix as a control before adding the
355enzyme. 1 µl of RNAse R was added to the remaining mix and solution was incubated at 37° C and
3565 µl were collected after 30 minutes, 1 hour and 2 hours. After each collection, RNAse R was
357inactivated through 5 minutes incubation at 80 °C. All the different timepoints of the reactions were
358analyzed by northern blot as described above.
359Phylogenetic analysis
360Phylogenetic analysis was performed on Cryphonectria parasitica mymonavirus 1 RdRP sequence
361to describe phylogenetic relations of the virus with those already characterized and available in the
362databases. ORF prediction on the viral contig was performed with ORFinder and predicted RdRP
363sequence was blasted to retrieve the closest viral sequence in the database. After choosing the
364RdRPs to include in the analysis, sequence alignment was performed using MUSCLE implemented

21 11
22
365in MEGA 6.0 (Edgar, 2004). Obtained alignment was exported in fasta format and submitted to
366IqTREE webserver (Trifinopoulos et al., 2016) to build the phylogenetic tree using ML method
367and performing automatic model selection. Ultrafast bootstrap analysis was performed by IqTREE
368with 1000 replicates. Branches under a bootstrap value of 50 were collapsed to polytomy .
369Sequences used for the phylogenetic analysis are shown in Suppl. table 4.
370

371Figure Legends
372Figure 1. Location of sampling districts in Azerbaijan that in 2018 and 2019 were assessed for the
373presence of chestnut blight hypovirus: Zagatala (1), Qakh (3), Shaki (2), Qabala (3).
374

375

376Literature quoted
377
378Aghayeva D., Harrington T. (2007) First report of Cryphonectria parasitica on chestnut (Castanea sativa) in
379 Azerbaijan. New Dis. Rep, 15, 33.
380Aghayeva D. N., Rigling D., Prospero S. (2017) Low genetic diversity but frequent sexual reproduction of the
381 chestnut blight fungus Cryphonectria parasitica in Azerbaijan. Forest Pathology, 47.
382Bertran A., Ciuffo M., Margaria P., Rosa C., Resende R. O., Turina M. (2016) Host-specific accumulation and
383 temperature effects on the generation of dimeric viral RNA species derived from the S-RNA of
384 members of the Tospovirus genus. Journal of General Virology, 97, 3051-3062.
385Buchfink B., Xie C., Huson D. H. (2015) Fast and sensitive protein alignment using DIAMOND. Nature
386 Methods, 12, 59-60.
387Crouch J. A., Dawe A., Aerts A., Barry K., Churchill A. C. L., Grimwood J., Hillman B., Milgroom M. G.,
388 Pangilinan J., Smith M., Salamov A., Schmutz J., Yadav J., Grigoriev I. V., Nuss D. (2020) Genome
389 sequence of the chestnut blight fungus Cryphonectria parasitica EP155: A fundamental resource for
390 an archetypical invasive plant pathogen. Phytopathology.
391Dawe A. L., Nuss D. L. (2013) Hypovirus Molecular Biology: From Koch's Postulates to Host Self-Recognition
392 Genes that Restrict Virus Transmission. Advances in Virus Research, Vol 86: Mycoviruses, 86, 109-
393 147.
394Deakin G., Dobbs E., Bennett J. M., Jones I. M., Grogan H. M., Burton K. S. (2017) Multiple viral infections in
395 Agaricus bisporus - Characterisation of 18 unique RNA viruses and 8 ORFans identified by deep
396 sequencing. Scientific Reports, 7.
397Edgar R. C. (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic
398 Acids Research, 32, 1792-1797.
399Eusebio-Cope A., Sun L. Y., Tanaka T., Chiba S., Kasahara S., Suzuki N. (2015) The chestnut blight fungus for
400 studies on virus/host and virus/virus interactions: From a natural to a model host. Virology, 477,
401 164-175.
402Ferriol I., Vallino M., Ciuffo M., Nigg J. C., Zamora-Macorra E. J., Falk B. W., Turina M. (2018) The
403 Torradovirus-specific RNA2-ORF1 protein is necessary for plant systemic infection. Molecular Plant
404 Pathology, 19, 1319-1331.
405Grente J. (1965) Les formes hypovirulentes d'Endothia parasitica et les espoirs de lutte contre le chancre du
406 chataignier. CR Acad. Agric. France, 51, 1033-1037.
407Haas B. J., Papanicolaou A., Yassour M., Grabherr M., Blood P. D., Bowden J., Couger M. B., Eccles D., Li B.,
408 Lieber M. (2013) De novo transcript sequence reconstruction from RNA-seq using the Trinity
409 platform for reference generation and analysis. Nature protocols, 8, 1494-1512.

23 12
24
410Hillman B. I., Suzuki N. (2004) Viruses of the chestnut blight fungus, Cryphonectria parasitica. Advances in
411 Virus Research, Vol. 63, 63, 423-472.
412Langmead B., Salzberg S. L. (2012) Fast gapped-read alignment with Bowtie 2. Nature Methods, 9, 357-
413 U354.
414Li H., Handsaker B., Wysoker A., Fennell T., Ruan J., Homer N., Marth G., Abecasis G., Durbin R., Genome
415 Project Data P. (2009) The Sequence Alignment/Map format and SAMtools. Bioinformatics, 25,
416 2078-2079.
417Liu L., Xie J., Cheng J., Fu Y., Li G., Yi X., Jiang D. (2014) Fungal negative-stranded RNA virus that is related to
418 bornaviruses and nyaviruses. Proceedings of the National Academy of Sciences of the United States
419 of America, 111, 12205-12210.
420Milgroom M. G., Cortesi P. (2004) Biological control of chestnut blight with hypovirulence: A critical
421 analysis. Annual Review of Phytopathology, 42, 311-338.
422Milgroom M. G., Lipari S. E. (1995) SPATIAL-ANALYSIS OF NUCLEAR AND MITOCHONDRIAL RFLP GENOTYPES
423 IN POPULATIONS OF THE CHESTNUT BLIGHT FUNGUS, CRYPHONECTRIA-PARASITICA. Molecular
424 Ecology, 4, 633-642.
425Milne I., Bayer M., Stephen G., Cardle L., Marshall D. (2016) Tablet: visualizing next-generation sequence
426 assemblies and mappings. In: Plant Bioinformatics, pp. 253-268. Springer.
427Nerva L., Forgia M., Ciuffo M., Chitarra W., Chiapello M., Vallino M., Varese G. C., Turina M. (2019) The
428 mycovirome of a fungal collection from the sea cucumber Holothuria polii. Virus research, 273,
429 197737-197737.
430Nerva L., Silvestri A., Ciuffo M., Palmano S., Varese G. C., Turina M. (2017) Transmission of Penicillium
431 aurantiogriseum partiti-like virus 1 to a new fungal host (Cryphonectria parasitica) confers higher
432 resistance to salinity and reveals adaptive genomic changes. Environmental Microbiology, 19, 4480-
433 4492.
434Nerva L., Varese G. C., Turina M. (2018) Different Approaches to Discover Mycovirus Associated to Marine
435 Organisms. Methods in molecular biology (Clifton, N.J.), 1746, 97-114.
436Nuss D. L. (2005) Hypovirulence: Mycoviruses at the fungal-plant interface. Nature Reviews Microbiology, 3,
437 632-642.
438Picarelli M., Forgia M., Rivas E. B., Nerva L., Chiapello M., Turina M., Colariccio A. (2019a) Extreme Diversity
439 of Mycoviruses Present in Isolates of Rhizoctonia solani AG2-2 LP From Zoysia japonica From Brazil.
440 Frontiers in Cellular and Infection Microbiology, 9.
441Picarelli M. A. S., Forgia M., Rivas E. B., Nerva L., Chiapello M., Turina M., Colariccio A. (2019b) Extreme
442 diversity of mycoviruses present in isolates of Rhizoctonia solani AG2-2 LP from Zoysia japonica
443 from Brazil. Frontiers in cellular and infection microbiology, 9, 244.
444Prospero S., Lutz A., Tavadze B., Supatashvili A., Rigling D. (2013) Discovery of a new gene pool and a high
445 genetic diversity of the chestnut blight fungus Cryphonectria parasitica in Caucasian Georgia.
446 Infection Genetics and Evolution, 20, 131-139.
447Puhalla J. E., Anagnostakis S. L. (1971) Genetic and nutritional requirements of Endothia parasitica.
448 Phytopathology, 61, 169-173.
449Rastgou M., Habibi M. K., Izadpanah K., Masenga V., Milne R. G., Wolf Y. I., Koonin E. V., Turina M. (2009)
450 Molecular characterization of the plant virus genus Ourmiavirus and evidence of inter-kingdom
451 reassortment of viral genome segments as its possible route of origin. Journal of General Virology,
452 90, 2525-2535.
453Rigling D., Prospero S. (2017) Cryphonectria parasitica, the causal agent of chestnut blight: Invasion history,
454 population biology and disease control. Molecular Plant Pathology.
455Salem N., Mansour A., Ciuffo M., Falk B. W., Turina M. (2016) A new tobamovirus infecting tomato crops in
456 Jordan. Archives of virology, 161, 503-506.
457Trifinopoulos J., Lam-Tung N., von Haeseler A., Minh B. Q. (2016) W-IQ-TREE: a fast online phylogenetic tool
458 for maximum likelihood analysis. Nucleic Acids Research, 44, W232-W235.
459Untergasser A., Cutcutache I., Koressaar T., Ye J., Faircloth B. C., Remm M., Rozen S. G. (2012) Primer3-new
460 capabilities and interfaces. Nucleic Acids Research, 40.

25 13
26
461

27 14
28